|
{ |
|
"best_metric": 0.7839844036299938, |
|
"best_model_checkpoint": "legal-bert-base-uncased/checkpoint-3858", |
|
"epoch": 14.0, |
|
"eval_steps": 500, |
|
"global_step": 9002, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.78, |
|
"grad_norm": 6.735908031463623, |
|
"learning_rate": 4.742353551062727e-05, |
|
"loss": 1.1096, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_accuracy": 0.7978311386522076, |
|
"eval_f1_macro": 0.613418966848975, |
|
"eval_f1_micro": 0.7978311386522076, |
|
"eval_f1_weighted": 0.7839880886224406, |
|
"eval_loss": 0.6747979521751404, |
|
"eval_macro_fpr": 0.018830502815301575, |
|
"eval_macro_sensitivity": 0.6339578261148517, |
|
"eval_macro_specificity": 0.9845345883924012, |
|
"eval_precision": 0.7854860044786733, |
|
"eval_precision_macro": 0.6239440349375213, |
|
"eval_recall": 0.7978311386522076, |
|
"eval_recall_macro": 0.6339578261148517, |
|
"eval_runtime": 30.9921, |
|
"eval_samples_per_second": 41.656, |
|
"eval_steps_per_second": 5.227, |
|
"eval_weighted_fpr": 0.017778080512226686, |
|
"eval_weighted_sensitivity": 0.7978311386522076, |
|
"eval_weighted_specificity": 0.9701876872338092, |
|
"step": 643 |
|
}, |
|
{ |
|
"epoch": 1.56, |
|
"grad_norm": 6.824150562286377, |
|
"learning_rate": 4.48315189217211e-05, |
|
"loss": 0.6187, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_accuracy": 0.8109992254066615, |
|
"eval_f1_macro": 0.7268228216676939, |
|
"eval_f1_micro": 0.8109992254066615, |
|
"eval_f1_weighted": 0.8089875323563979, |
|
"eval_loss": 0.6449354887008667, |
|
"eval_macro_fpr": 0.016938802987367406, |
|
"eval_macro_sensitivity": 0.7327307314578203, |
|
"eval_macro_specificity": 0.9857686909763926, |
|
"eval_precision": 0.8196084470842065, |
|
"eval_precision_macro": 0.7805936449528967, |
|
"eval_recall": 0.8109992254066615, |
|
"eval_recall_macro": 0.7327307314578203, |
|
"eval_runtime": 58.3071, |
|
"eval_samples_per_second": 22.141, |
|
"eval_steps_per_second": 2.778, |
|
"eval_weighted_fpr": 0.016373641121997046, |
|
"eval_weighted_sensitivity": 0.8109992254066615, |
|
"eval_weighted_specificity": 0.9755311392392266, |
|
"step": 1286 |
|
}, |
|
{ |
|
"epoch": 2.33, |
|
"grad_norm": 20.75279998779297, |
|
"learning_rate": 4.223950233281493e-05, |
|
"loss": 0.4747, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"eval_accuracy": 0.814872192099148, |
|
"eval_f1_macro": 0.7369957890904685, |
|
"eval_f1_micro": 0.814872192099148, |
|
"eval_f1_weighted": 0.8124596447054482, |
|
"eval_loss": 0.8151384592056274, |
|
"eval_macro_fpr": 0.016571707879599946, |
|
"eval_macro_sensitivity": 0.739039840019417, |
|
"eval_macro_specificity": 0.9860629698102752, |
|
"eval_precision": 0.8192276160546627, |
|
"eval_precision_macro": 0.7659183928188651, |
|
"eval_recall": 0.814872192099148, |
|
"eval_recall_macro": 0.739039840019417, |
|
"eval_runtime": 55.9094, |
|
"eval_samples_per_second": 23.091, |
|
"eval_steps_per_second": 2.898, |
|
"eval_weighted_fpr": 0.015968463954032203, |
|
"eval_weighted_sensitivity": 0.814872192099148, |
|
"eval_weighted_specificity": 0.9760723550549795, |
|
"step": 1929 |
|
}, |
|
{ |
|
"epoch": 3.11, |
|
"grad_norm": 10.737870216369629, |
|
"learning_rate": 3.9652669777086575e-05, |
|
"loss": 0.3866, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 3.89, |
|
"grad_norm": 13.75886344909668, |
|
"learning_rate": 3.7060653188180404e-05, |
|
"loss": 0.2645, |
|
"step": 2500 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_accuracy": 0.8218435321456236, |
|
"eval_f1_macro": 0.738478964900184, |
|
"eval_f1_micro": 0.8218435321456236, |
|
"eval_f1_weighted": 0.8189030195253059, |
|
"eval_loss": 0.9344585537910461, |
|
"eval_macro_fpr": 0.015763618193954002, |
|
"eval_macro_sensitivity": 0.7412892486668318, |
|
"eval_macro_specificity": 0.986618111597957, |
|
"eval_precision": 0.8198244063808544, |
|
"eval_precision_macro": 0.7446229748697999, |
|
"eval_recall": 0.8218435321456236, |
|
"eval_recall_macro": 0.7412892486668318, |
|
"eval_runtime": 75.578, |
|
"eval_samples_per_second": 17.082, |
|
"eval_steps_per_second": 2.143, |
|
"eval_weighted_fpr": 0.015247944842216919, |
|
"eval_weighted_sensitivity": 0.8218435321456236, |
|
"eval_weighted_specificity": 0.9774281418237326, |
|
"step": 2572 |
|
}, |
|
{ |
|
"epoch": 4.67, |
|
"grad_norm": 8.32991886138916, |
|
"learning_rate": 3.4468636599274234e-05, |
|
"loss": 0.1901, |
|
"step": 3000 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"eval_accuracy": 0.8195197521301317, |
|
"eval_f1_macro": 0.7594855831156184, |
|
"eval_f1_micro": 0.8195197521301317, |
|
"eval_f1_weighted": 0.81658937168797, |
|
"eval_loss": 1.0929232835769653, |
|
"eval_macro_fpr": 0.01614879224826684, |
|
"eval_macro_sensitivity": 0.743232314276478, |
|
"eval_macro_specificity": 0.9863035723232563, |
|
"eval_precision": 0.8241924882578779, |
|
"eval_precision_macro": 0.8264084869739744, |
|
"eval_recall": 0.8195197521301317, |
|
"eval_recall_macro": 0.743232314276478, |
|
"eval_runtime": 65.7326, |
|
"eval_samples_per_second": 19.64, |
|
"eval_steps_per_second": 2.465, |
|
"eval_weighted_fpr": 0.01548687271518777, |
|
"eval_weighted_sensitivity": 0.8195197521301317, |
|
"eval_weighted_specificity": 0.9750338327187125, |
|
"step": 3215 |
|
}, |
|
{ |
|
"epoch": 5.44, |
|
"grad_norm": 50.74209976196289, |
|
"learning_rate": 3.187662001036807e-05, |
|
"loss": 0.1131, |
|
"step": 3500 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"eval_accuracy": 0.820294345468629, |
|
"eval_f1_macro": 0.7839844036299938, |
|
"eval_f1_micro": 0.8202943454686291, |
|
"eval_f1_weighted": 0.8197212343639901, |
|
"eval_loss": 1.1536158323287964, |
|
"eval_macro_fpr": 0.01593810671400651, |
|
"eval_macro_sensitivity": 0.7786487516932519, |
|
"eval_macro_specificity": 0.9864617261648754, |
|
"eval_precision": 0.821246410485655, |
|
"eval_precision_macro": 0.7968460348821496, |
|
"eval_recall": 0.820294345468629, |
|
"eval_recall_macro": 0.7786487516932519, |
|
"eval_runtime": 54.9389, |
|
"eval_samples_per_second": 23.499, |
|
"eval_steps_per_second": 2.949, |
|
"eval_weighted_fpr": 0.015407092575375215, |
|
"eval_weighted_sensitivity": 0.820294345468629, |
|
"eval_weighted_specificity": 0.9766315470045015, |
|
"step": 3858 |
|
}, |
|
{ |
|
"epoch": 6.22, |
|
"grad_norm": 0.02055634744465351, |
|
"learning_rate": 2.928978745463971e-05, |
|
"loss": 0.0975, |
|
"step": 4000 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"grad_norm": 32.151573181152344, |
|
"learning_rate": 2.669777086573354e-05, |
|
"loss": 0.063, |
|
"step": 4500 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"eval_accuracy": 0.8117738187451587, |
|
"eval_f1_macro": 0.7495479890901372, |
|
"eval_f1_micro": 0.8117738187451587, |
|
"eval_f1_weighted": 0.8135965211162124, |
|
"eval_loss": 1.3217942714691162, |
|
"eval_macro_fpr": 0.01664162427116977, |
|
"eval_macro_sensitivity": 0.7526435450791772, |
|
"eval_macro_specificity": 0.9859379584321952, |
|
"eval_precision": 0.818420676521013, |
|
"eval_precision_macro": 0.7517839435012339, |
|
"eval_recall": 0.8117738187451587, |
|
"eval_recall_macro": 0.7526435450791772, |
|
"eval_runtime": 59.9636, |
|
"eval_samples_per_second": 21.53, |
|
"eval_steps_per_second": 2.702, |
|
"eval_weighted_fpr": 0.016292323164599398, |
|
"eval_weighted_sensitivity": 0.8117738187451587, |
|
"eval_weighted_specificity": 0.9772955577377688, |
|
"step": 4501 |
|
}, |
|
{ |
|
"epoch": 7.78, |
|
"grad_norm": 0.0045046876184642315, |
|
"learning_rate": 2.4105754276827372e-05, |
|
"loss": 0.0264, |
|
"step": 5000 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"eval_accuracy": 0.82571649883811, |
|
"eval_f1_macro": 0.7730278469427766, |
|
"eval_f1_micro": 0.82571649883811, |
|
"eval_f1_weighted": 0.8247069887539016, |
|
"eval_loss": 1.3862823247909546, |
|
"eval_macro_fpr": 0.015458317393184977, |
|
"eval_macro_sensitivity": 0.7767515241074096, |
|
"eval_macro_specificity": 0.9868314420976501, |
|
"eval_precision": 0.8261549954416628, |
|
"eval_precision_macro": 0.7783699125562222, |
|
"eval_recall": 0.82571649883811, |
|
"eval_recall_macro": 0.7767515241074096, |
|
"eval_runtime": 58.9779, |
|
"eval_samples_per_second": 21.89, |
|
"eval_steps_per_second": 2.747, |
|
"eval_weighted_fpr": 0.01485246550927454, |
|
"eval_weighted_sensitivity": 0.82571649883811, |
|
"eval_weighted_specificity": 0.97675513262664, |
|
"step": 5144 |
|
}, |
|
{ |
|
"epoch": 8.55, |
|
"grad_norm": 0.007681012153625488, |
|
"learning_rate": 2.1513737687921205e-05, |
|
"loss": 0.03, |
|
"step": 5500 |
|
}, |
|
{ |
|
"epoch": 9.0, |
|
"eval_accuracy": 0.8079008520526724, |
|
"eval_f1_macro": 0.7594758314568825, |
|
"eval_f1_micro": 0.8079008520526724, |
|
"eval_f1_weighted": 0.8095900464774608, |
|
"eval_loss": 1.5542314052581787, |
|
"eval_macro_fpr": 0.017245749289239998, |
|
"eval_macro_sensitivity": 0.7652642279270354, |
|
"eval_macro_specificity": 0.9854886624123184, |
|
"eval_precision": 0.8166769511091512, |
|
"eval_precision_macro": 0.7639132572736053, |
|
"eval_recall": 0.8079008520526724, |
|
"eval_recall_macro": 0.7652642279270354, |
|
"eval_runtime": 70.914, |
|
"eval_samples_per_second": 18.205, |
|
"eval_steps_per_second": 2.284, |
|
"eval_weighted_fpr": 0.0167003367003367, |
|
"eval_weighted_sensitivity": 0.8079008520526724, |
|
"eval_weighted_specificity": 0.9744290841320995, |
|
"step": 5787 |
|
}, |
|
{ |
|
"epoch": 9.33, |
|
"grad_norm": 0.0017093586502596736, |
|
"learning_rate": 1.8926905132192844e-05, |
|
"loss": 0.0149, |
|
"step": 6000 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"eval_accuracy": 0.8140975987606507, |
|
"eval_f1_macro": 0.7412163056128125, |
|
"eval_f1_micro": 0.8140975987606507, |
|
"eval_f1_weighted": 0.8126789471820407, |
|
"eval_loss": 1.583540678024292, |
|
"eval_macro_fpr": 0.016780932672955, |
|
"eval_macro_sensitivity": 0.7360563323248509, |
|
"eval_macro_specificity": 0.9858047486397965, |
|
"eval_precision": 0.8155262118469742, |
|
"eval_precision_macro": 0.7545069126135399, |
|
"eval_recall": 0.8140975987606507, |
|
"eval_recall_macro": 0.7360563323248509, |
|
"eval_runtime": 70.4882, |
|
"eval_samples_per_second": 18.315, |
|
"eval_steps_per_second": 2.298, |
|
"eval_weighted_fpr": 0.016049217600641968, |
|
"eval_weighted_sensitivity": 0.8140975987606507, |
|
"eval_weighted_specificity": 0.9729736308362953, |
|
"step": 6430 |
|
}, |
|
{ |
|
"epoch": 10.11, |
|
"grad_norm": 0.004950440023094416, |
|
"learning_rate": 1.6334888543286677e-05, |
|
"loss": 0.0092, |
|
"step": 6500 |
|
}, |
|
{ |
|
"epoch": 10.89, |
|
"grad_norm": 0.004135847091674805, |
|
"learning_rate": 1.374287195438051e-05, |
|
"loss": 0.005, |
|
"step": 7000 |
|
}, |
|
{ |
|
"epoch": 11.0, |
|
"eval_accuracy": 0.8241673121611154, |
|
"eval_f1_macro": 0.7681252828650814, |
|
"eval_f1_micro": 0.8241673121611154, |
|
"eval_f1_weighted": 0.8226392641142015, |
|
"eval_loss": 1.5324584245681763, |
|
"eval_macro_fpr": 0.015637185816662357, |
|
"eval_macro_sensitivity": 0.7811775190571517, |
|
"eval_macro_specificity": 0.9866649056306106, |
|
"eval_precision": 0.8249899677721785, |
|
"eval_precision_macro": 0.780545510593793, |
|
"eval_recall": 0.8241673121611154, |
|
"eval_recall_macro": 0.7811775190571517, |
|
"eval_runtime": 67.2418, |
|
"eval_samples_per_second": 19.199, |
|
"eval_steps_per_second": 2.409, |
|
"eval_weighted_fpr": 0.015010249289162203, |
|
"eval_weighted_sensitivity": 0.8241673121611154, |
|
"eval_weighted_specificity": 0.9758062722980435, |
|
"step": 7073 |
|
}, |
|
{ |
|
"epoch": 11.66, |
|
"grad_norm": 0.0040180375799536705, |
|
"learning_rate": 1.115085536547434e-05, |
|
"loss": 0.003, |
|
"step": 7500 |
|
}, |
|
{ |
|
"epoch": 12.0, |
|
"eval_accuracy": 0.8288148721920992, |
|
"eval_f1_macro": 0.7625723813018138, |
|
"eval_f1_micro": 0.8288148721920992, |
|
"eval_f1_weighted": 0.8276081900441157, |
|
"eval_loss": 1.5713844299316406, |
|
"eval_macro_fpr": 0.015190042544950861, |
|
"eval_macro_sensitivity": 0.7678635910180709, |
|
"eval_macro_specificity": 0.9870237726059469, |
|
"eval_precision": 0.8299388397158192, |
|
"eval_precision_macro": 0.7700952459098459, |
|
"eval_recall": 0.8288148721920992, |
|
"eval_recall_macro": 0.7678635910180709, |
|
"eval_runtime": 61.0459, |
|
"eval_samples_per_second": 21.148, |
|
"eval_steps_per_second": 2.654, |
|
"eval_weighted_fpr": 0.014538517202815604, |
|
"eval_weighted_sensitivity": 0.8288148721920992, |
|
"eval_weighted_specificity": 0.9765417168971026, |
|
"step": 7716 |
|
}, |
|
{ |
|
"epoch": 12.44, |
|
"grad_norm": 0.004260718822479248, |
|
"learning_rate": 8.55883877656817e-06, |
|
"loss": 0.0033, |
|
"step": 8000 |
|
}, |
|
{ |
|
"epoch": 13.0, |
|
"eval_accuracy": 0.8249419054996127, |
|
"eval_f1_macro": 0.7607533001144507, |
|
"eval_f1_micro": 0.8249419054996127, |
|
"eval_f1_weighted": 0.8224643561391238, |
|
"eval_loss": 1.551088571548462, |
|
"eval_macro_fpr": 0.015560628647913471, |
|
"eval_macro_sensitivity": 0.7597662805173221, |
|
"eval_macro_specificity": 0.9867279232222673, |
|
"eval_precision": 0.821925563604887, |
|
"eval_precision_macro": 0.767644957538908, |
|
"eval_recall": 0.8249419054996127, |
|
"eval_recall_macro": 0.7597662805173221, |
|
"eval_runtime": 66.776, |
|
"eval_samples_per_second": 19.333, |
|
"eval_steps_per_second": 2.426, |
|
"eval_weighted_fpr": 0.014931289640591967, |
|
"eval_weighted_sensitivity": 0.8249419054996127, |
|
"eval_weighted_specificity": 0.9759769428343987, |
|
"step": 8359 |
|
}, |
|
{ |
|
"epoch": 13.22, |
|
"grad_norm": 0.001955605112016201, |
|
"learning_rate": 5.9668221876620015e-06, |
|
"loss": 0.0043, |
|
"step": 8500 |
|
}, |
|
{ |
|
"epoch": 14.0, |
|
"grad_norm": 0.0009153374703601003, |
|
"learning_rate": 3.379989631933644e-06, |
|
"loss": 0.0018, |
|
"step": 9000 |
|
}, |
|
{ |
|
"epoch": 14.0, |
|
"eval_accuracy": 0.8249419054996127, |
|
"eval_f1_macro": 0.7571714820627711, |
|
"eval_f1_micro": 0.8249419054996127, |
|
"eval_f1_weighted": 0.8224001833855348, |
|
"eval_loss": 1.5509757995605469, |
|
"eval_macro_fpr": 0.015514116711070718, |
|
"eval_macro_sensitivity": 0.7554168858318507, |
|
"eval_macro_specificity": 0.9867787829147775, |
|
"eval_precision": 0.8225003050604829, |
|
"eval_precision_macro": 0.768615439122566, |
|
"eval_recall": 0.8249419054996127, |
|
"eval_recall_macro": 0.7554168858318507, |
|
"eval_runtime": 56.3702, |
|
"eval_samples_per_second": 22.902, |
|
"eval_steps_per_second": 2.874, |
|
"eval_weighted_fpr": 0.014931289640591967, |
|
"eval_weighted_sensitivity": 0.8249419054996127, |
|
"eval_weighted_specificity": 0.97673983822205, |
|
"step": 9002 |
|
} |
|
], |
|
"logging_steps": 500, |
|
"max_steps": 9645, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 15, |
|
"save_steps": 500, |
|
"total_flos": 1.892831352227021e+16, |
|
"train_batch_size": 8, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|